Closed pkofod closed 8 years ago
Ref: https://gist.github.com/anonymous/4609c24a5c7b8bb8222012c9564958c6 and https://github.com/JuliaLang/METADATA.jl/pull/5341
The way traces are initiated has changed a bit in Optim. I didn't notice you were accessing OptimizationTrace directly, so currently Mads.jl tests will fail in contamination.jl with Optim v0.5, because a trace cannot be initiated.
OptimizationTrace
Mads.jl
contamination.jl
Does this work with any earlier versions of Optim? If not, then best to set a lower bound of Optim 0.5 now for this change.
Optim 0.5
Good point, changed.
Changes Unknown when pulling bdc7f27a66cf789204b0bd5cf07355d27d8e1b55 on pkofod:master into \ on madsjulia:master**.
Ref: https://gist.github.com/anonymous/4609c24a5c7b8bb8222012c9564958c6 and https://github.com/JuliaLang/METADATA.jl/pull/5341
The way traces are initiated has changed a bit in Optim. I didn't notice you were accessing
OptimizationTrace
directly, so currentlyMads.jl
tests will fail incontamination.jl
with Optim v0.5, because a trace cannot be initiated.